2D Cavity Flow Simulation only depends on python numpy The Source code and techniques are modified from https://github.com/mathworks/2D-Lid-Driven-Cavity-Flow-Incompressible-Navier-Stokes-Solver.git
This work is a re-implement of matlab work.
Anything about solving linear equations is replaced by numpy FFT solver
And the stem of this work can be transplanted into cupy module.
The 'Stable_CavityFlow_2or3order.py' gives the Crank-Nicolson(2th order) and RK3(3th order)
version in the original manuscripts while 'Euler_forward.py' gives a 1th order
numerical formulation.
The simulation streamlines are as follows:
A cupy version of data parallel version is released in gpuCavityFlow.py The memory of Nvidia-gpu can be exploited efficiently by solving thousands of 2d equations simultaneously.
A numpy motivated Cylinder's winding flow has been submitted in cylinder.mp4;
And the figure can be seen in
Now a 3d-cavity simulation with pytorch implement is released and
the visualization can be completed on Matlab or paraview, the images
derived from 知乎
Also a cahn-hilliard-FDM solver has been committed compromising 2d and 3d implements
